From 18fd0c778533c168e89be212426811229ebf1a63 Mon Sep 17 00:00:00 2001
From: Rahul Narla <rahul.narla@commure.com>
Date: Tue, 24 Feb 2026 12:22:09 -0800
Subject: [PATCH] script: pull vocab locally

---
 .gitignore                 |  5 ++++-
 scripts/sync-vocabulary.sh | 30 ++++++++++++++++++++++++++++++
 2 files changed, 34 insertions(+), 1 deletion(-)
 create mode 100755 scripts/sync-vocabulary.sh

diff --git a/.gitignore b/.gitignore
index 1a533534..ca019c35 100644
--- a/.gitignore
+++ b/.gitignore
@@ -7,4 +7,7 @@
 .classpath
 /src/main/profiles/local/
 /bin
-/reference-ccda-api
\ No newline at end of file
+/reference-ccda-api
+
+# Validator data (configs, vocabulary, scenarios) – populated by scripts or locally
+/validator/
\ No newline at end of file
diff --git a/scripts/sync-vocabulary.sh b/scripts/sync-vocabulary.sh
new file mode 100755
index 00000000..2b54655f
--- /dev/null
+++ b/scripts/sync-vocabulary.sh
@@ -0,0 +1,30 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# Sync vocabulary data from GCS into validator/vocabulary.
+# Requires: gcloud CLI installed and authenticated with access to the bucket.
+#
+# Usage:
+#   ./scripts/sync-vocabulary.sh
+#
+# First-time auth (if needed):
+#   gcloud auth login
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+VOCAB_DIR="$REPO_ROOT/validator/vocabulary"
+GCS_VOCAB="gs://health-bridge-st-ccda-validator-vocabulary/vocabulary"
+
+echo "Ensuring gcloud is authenticated..."
+if ! gcloud auth print-access-token &>/dev/null; then
+  echo "Run: gcloud auth login"
+  exit 1
+fi
+
+echo "Creating local directory: $VOCAB_DIR"
+mkdir -p "$VOCAB_DIR"
+
+echo "Syncing from $GCS_VOCAB to $VOCAB_DIR ..."
+gsutil -m rsync -r "$GCS_VOCAB" "$VOCAB_DIR"
+
+echo "Done. Vocabulary is in validator/vocabulary (ignored by .gitignore)."